Core i7 Overclocking by Batch #

[Erdei88]

3849A846

[imo]

[-Sweeper_]

mine is 3941B077

pretty curious to see how far it goes

[S1N3]

3941 B047, 21x200 4.2 stable at 1.4core 1.96 pll 1.35 qpi. Max temp 63c. In my opinion not an outstanding example.
All dimms occupied
max qpi link
HT enabled

[CedricFP]

3924A

21x200 @ 1.31v
+150mv VTT
1.35 PLL
1.3 QPI
1.7 vDIMM

PCIE @ 101

[aamsel]

3941 B047, 21x200 4.2 stable at 1.4core 1.96 pll 1.35 qpi. Max temp 63c. In my opinion not an outstanding example.
All dimms occupied
max qpi link
HT enabled

Just picked up a 3941B047 as well.
The seller (EBAY) claimed the following performance for this batch:

Current batch number is 3941B047
My testing results
2.66ghz @ 0.93V
3GHZ @ 0.944V
4.0ghz @ 1.21V
4.488mhz @ 1.425V
4.640mhz @ 1.425V (not stable)
My motherboard is Asus P6T deluxe with beta bios 0006
Memory G-skill perfect storm 2000mhz

Perhaps a slight exaggeration?

[Salamndar]

is it me or is itt really that increasing batch number makes CPUs less exiting, the first batches of D0 used to overclock better than current batches and uses less voltage!

I have a D0 3923A and I need 1.36v @ CPU, 1.22@ QPI/DRAM and 1.81@ PLL just to reach 4.2GHz, if I want to reach 4.4 which is the maximum I could achieve with my R2E then I have to ridiculously increasing voltage as following:
1.46@ CPU, 1.32 @ QPI/DRAM and 1.89@ PLL !!
That's with Blck 210, any thing higher than that and it is very instable at any voltage also any thing higher than 213Blck and the PC won't POST.
I have to add that IOH and ICH needs a bump in voltage also.

[aamsel]

is it me or is itt really that increasing batch number makes CPUs less exiting, the first batches of D0 used to overclock better than current batches and uses less voltage!

I have a D0 3923A and I need 1.36v @ CPU, 1.22@ QPI/DRAM and 1.81@ PLL just to reach 4.2GHz, if I want to reach 4.4 which is the maximum I could achieve with my R2E then I have to ridiculously increasing voltage as following:
1.46@ CPU, 1.32 @ QPI/DRAM and 1.89@ PLL !!
That's with Blck 210, any thing higher than that and it is very instable at any voltage also any thing higher than 213Blck and the PC won't POST.
I have to add that IOH and ICH needs a bump in voltage also.

I think that INTEL could be binning better i7's as Xeons more now
than they did early in D0, and that is why the 3849B, 3845B, etc. are the better overclockers. Basically, INTEL is probably far enough along into D0 Bloomfield process that they are better setup to keep the absolutely known primo chips from being sold as 920's.

The odd part is that some people say report that oc Xeon Bloomfiled
is the best, and some apparently not. Still luck of draw, no doubt, as CPU's are definitely unique.

Other than that, when can we expect to find batch "C" 920's?

[Salamndar]

Batch "C " !?!?!?

DO you mean XXXXCXXX ??

[ChaosAD]

I got a 3941B092. Since i m on stock intel cooler atm , the best i could achieve is 21x182 @ 1.20v (3 passes of LinX, stopped after cause of 87C on core).

[Salamndar]

I got a 3941B092. Since i m on stock intel cooler atm , the best i could achieve is 21x182 @ 1.20v (3 passes of LinX, stopped after cause of 87C on core).

I wouldn't even consider overclock 100 MHz over stock speed with a stock fan.
You Mr. have some balls !!!

Last month before I spent the money had on an HTC Hero mobile phone, I was tempted to buy either a 3932Bxxx 920 or a 393xA xeon 3520.

From your experience guys, which one is better, Please note that there is a price def. 100$ for the xeon.

[aamsel]

Batch "C " !?!?!?

DO you mean XXXXCXXX ??

Yes, but no idea if it exists or will exist.
At this point, it is just talk.

[SpinCykle]

Picked up one the other day for 200$ at a local MC. Cant find any info on the FPO/Batch#. Maybe you guys can help! I'm still waiting on MB, RAM and blocks before I get everything up and running.

Core i7 920 D0
SLBEJ
FPO/Batch# 3918A495

[Salamndar]

OK guys, I'm in the market to buy me a new single slot Geforce to use it as a physX with my CrossFireX 4870X2 and I came across these :

a Xeon for 440$ :



and a 920 for 320 $:



Is the 120$ worth it ? I'm not sure really.
Also any word over these batches ?

[kup]

I'd go for the 3918B083 stepped chip.

[Salamndar]

I have the 3923A batch and I'm not sure is my R2E limiting me to 210 Bclk or the CPU doesn't like any thing higher than 210, I don't want to burn 320$ just to find out it is a mobo issue !

But again there is no way to tell other than switching the CPU!

[uranium]

3845B260 @ 4.881

[IMG][/IMG]

Finding the next 19MZ to get 4.9Ghz a slight challenge.....

[Salamndar]

Don't tell me that's on water !!?!?

[uranium]

Don't tell me that's on water !!?!?

Water chilled to 10 deg C Won't do that with HT on though

[Salamndar]

Nice indeed!

[uranium]

Nice indeed!

Thank you

[Clint]

3942A***** , does 4.0GHz on 1.2v Prime95 stable (2h) outside chassi and with a Scythe mini-ninja with two 1200rpm push/pull fans giving a max temp of 70c. Did take me a while to find out how to oc the i7 architechture, coming from core2. Once I got hang of qpi and pll importnace my voltage could be reduced from 1.35 to 1.2v. I'ts mounted on a MSI X598M board together with 6 Mushkin 1600MHz sticks. Board does 225 BCLK with mem slots fully occupied, very nice board for that price! Gonna drop a better cooler on it and se what we can do. Is there any need for lapping the i7 cpus? My old q9300 dropped 8c when giving it a overhaul with 1200 paper...

[rusty]

3942A***** , does 4.0GHz on 1.2v Prime95 stable (2h) outside chassi and with a Scythe mini-ninja with two 1200rpm push/pull fans giving a max temp of 70c. Did take me a while to find out how to oc the i7 architechture, coming from core2. Once I got hang of qpi and pll importnace my voltage could be reduced from 1.35 to 1.2v. I'ts mounted on a MSI X598M board together with 6 Mushkin 1600MHz sticks. Board does 225 BCLK with mem slots fully occupied, very nice board for that price! Gonna drop a better cooler on it and se what we can do. Is there any need for lapping the i7 cpus? My old q9300 dropped 8c when giving it a overhaul with 1200 paper...

what's your qpi and pll volts now ?

[ChaosAD]

3942A***** , does 4.0GHz on 1.2v Prime95 stable (2h) outside chassi and with a Scythe mini-ninja with two 1200rpm push/pull fans giving a max temp of 70c. Did take me a while to find out how to oc the i7 architechture, coming from core2. Once I got hang of qpi and pll importnace my voltage could be reduced from 1.35 to 1.2v. I'ts mounted on a MSI X598M board together with 6 Mushkin 1600MHz sticks. Board does 225 BCLK with mem slots fully occupied, very nice board for that price! Gonna drop a better cooler on it and se what we can do. Is there any need for lapping the i7 cpus? My old q9300 dropped 8c when giving it a overhaul with 1200 paper...

Plz post the changes that u made to drop vcore. I would also like to know.

[Clint]

Plz post the changes that u made to drop vcore. I would also like to know.

Well, I did use the auto option in MSI bios to help me get up to 4.0-4.2 on max auto volts which was 1.365v. Auto feature gave me a max 1.7v on pll and 1.39 on qpi. In a perfect world the auto function would know the specific character of my CPU, but I guess they rather trow in a bit too much on the qpi when going on high bclk and staying "safe" on the pll (wich are rumoured to kill CPU's). I did some quick research and lowered the qpi to be a bit closer to my intended vcore of 1.2v and landed on 1.35v. As for pll even Intel states that 1.8v is fine, so I changed that to 1.8v. Then I just throw a session of Prime95 and started to lower the vcore in windows with the MSI oc util. Landed on 1.2v, you can imagine my joy when I continued to lower the voltage and the primesession kept on going...

4GHz on 1.2v is not supergood, but it sure is a helova lot betta than 4GHz on 1.365v... Just bought a True Black and gonna lap that and while I'm at it lap the CPU as well. Just for fun...